Crashes, red flags and restarts break the continuity of a race, and with it the continuity of the data. A chart that looks smooth can hide the fact that the session was split into parts by an interruption.
Interruptions and data
After a red flag, the tyres are changed, the track may have changed, and the order is reset. Lap times before and after cannot be put on one line without a gap. If part of the race is cancelled or shortened, even the number of laps may differ from the original plan.
A good timing page, such as the Bet Rolla Casino, marks interruptions clearly, while a poor one lets the lines join up as if nothing had happened. Looking for the marker is a quick test of whether a source understands its own data.
Reading a crash fairly
A rider who crashes while pushing may be running a competitive pace, and one who crashes alone in practice has not necessarily revealed a weakness. Counting crashes as a single statistic hides these differences. Context from the session notes helps to separate carelessness from calculated risk.
Restarts raise the question of what counts as the same race. When a race is restarted, how the parts are scored depends on the regulations in force. A careful statistic names the rule it follows, and a careful reader looks for that note before comparing totals. Without it, two tables that look alike can measure different things. A practical test of a chart is to ask what happened at the break. When nothing is marked at the break, the chart is hiding something. A good chart notes the break, the tyre change and the new start order in a few words.
A red flag is a natural pause, so take one yourself. If betting markets are open, a budget and a time limit set in advance matter most when the situation is confusing.
As a general rule, any statistic that ignores interruptions overstates its precision. When a race is stopped, the honest approach is to analyse the segments separately and to say plainly where the sample is thin.
A broken record should be read in pieces. That supports understanding and avoids forecasts built on a line that never existed.